Hi Avinash: You could create a vector using c(), e.g., if the first 3 are control and last 3 are cancer, then, > myclassvec <- c(0,0,0,1,1,1) For creating the gene name matrix, I hope you have the probe annotation file in which each probe (or probe set, for affy) is listed along with gene name, etc. From that select index, genename, id and save these three columns to another tab-delimited file (named say, myfile.tab). Keep no header. Read this file using read.table, e.g., > mygnames <- read.table(file="myfile.tab", header=F, sep="\t") Have a look at ?read.table for more help. Good luck! Saroj avinash gupta wrote: > Saroj, > thank you, > class vec: my colum lenth is 6 so what is the code for make it . > genenames : i had download my data from in raw file and load .cel > file in R after that i normailze it and make the object for make the > matrix data i use tah command: > >expr<- exprs(obj) > after that what i do for creat the genenames matrix object, how to > creat 3 object and where to download the gene names and how to laos it > in genenames matrix object.plz mail me the code or manual for it . > > thank u > > > regards.. > > avinash

Hi Avinash So, it looks from your code like you haven't provided any phenotype information about your CEL files, and therefore each CEL file will be treated as a different condition. Because you have only 1 array per condition (i.e. you have no replicates), limma is going to give an error. Please note the following line from the puma User Guide (bottom of page 4 if you're using puma 2.0.0, but you still haven't given me the output from sessionInfo() so I don't know what version you're using): "The easiest way to supply phenotype information is in a text file that is loaded using the phenotype parameter of the ReadAffy function". Unfortunately, my documentation is incorrect here - ReadAffy has no phenotype parameter! Instead it should say phenoData parameter. As I suggested in my first reply, showing the output of pData(eset_rma) would confirm that the above is true. From reading ?ReadAffy, my guess is that pData(eset_rma) will give you a data.frame with a column called sample containing the numbers 1 to 6. To provide phenotype data for your eset_rma object you could do something like the following: pData(eset_rma) <- data.frame(AvinashCondition=c("Condition1", "Condition1", "Condition1", "Condition2", "Condition2", "Condition2")) "contrast" is a standard statistical term. I've tried to explain this in section 4.6 of the puma manual, but you could also look at the limma user guide, or google this (e.g. to get here: http://en.wikipedia.org/wiki/Contrast_%28statistics%29), or ask a local statistician. Please do read the posting guide - it can be found here: http://www.bioconductor.org/docs/postingGuide.html Best wishes Richard avinash gupta wrote: > sir, > puma clustering method . i make the object of (sample6, Celfile) > by this command > > dat<-ReadAffy() > > library(pumadata) > > >eset_mmgmos<-mmgmos(dat) > >eset_rma<-rma(dat) > > after that i do "dentifying differentially expressed genes" method in > this method i use this command > >eset_comb <- pumaComb(eset_mmgmos) > > pumaDERes <- pumaDE(eset_comb) > > limmaRes <- calculateLimma(eset_rma) ## in this command it show > the error like: > Error in ebayes(fit = fit, proportion = proportion, stdev.coef.lim = > stdev.coef.lim) : > No residual degrees of freedom in linear model fits > i don't understand it > plz solve this and mail me . > > i more things > in the code > toppumaDEIntGene<-topGenes(pumaDERes, contrast = 7) > > i don't under stand the value of contrast, how it's value define ,in my > obj have 6 cel file ,so plz help me to define its value. -- Dr Richard D Pearson richard.pearson at well.ox.ac.uk Wellcome Trust Centre for Human Genetics http://www.well.ox.ac.uk/~rpearson University of Oxford Tel: +44 (0)1865 617890 Roosevelt Drive Mob: +44 (0)7971 221181 Oxford OX3 7BN, UK Fax: +44 (0)1865 287664